Quick viewing(Text Mode)

Nproliferation Regime

Nproliferation Regime

MOHINI RAWOOL-SULLIVAN, PAUL D. MOSKOWITZ, & LUDMILA SHELENKOVA

Report

Technical and Proliferation-Related Aspects of the Dismantlement of Russian Alfa-Class Nuclear

MOHINI RAWOOL-SULLIVAN, PAUL D. MOSKOWITZ, & LUDMILA N. SHELENKOVA

Mohini Rawool-Sullivan is a technical staff member at the Los Alamos National Laboratory. Her areas of expertise include assessing nuclear threats and building radiation sensors for the nonproliferation regime. Paul D. Moskowitz is the manager of the Environmental Threat Reduction Program in the Nonproliferation and National Security Department at Brookhaven National Laboratory. He has been working on decommissioning and dismantlement projects for the U.S. Department of Defense, the U.S. Department of Energy and the U.S. Environmental Protection Agency since 1995. Dr. Ludmila N. Shelenkova is an assistant biophysicist at the Brookhaven National Laboratory, Nonproliferation and National Security Department. She develops, implements, and coordinates programs based on collaboration with Russian weapon scientists.

etween 1950 and 1994, the built a to the and one-third are assigned to the total of 245 nuclear submarines. At the end of Pacific Fleet. B1994, the newly formed Russian Federation (RF) Nuclear-powered submarines have provided the Soviet had the largest fleet of nuclear-powered submarines and and Russian navies with significant advantages over die- surface in the world. At that time, the Russian Fed- sel submarines. Nuclear submarines can stay submerged eration had a total of 140 active nuclear-powered vessels for months at a time and hence can patrol wide underwa- in its fleet. In addition to nuclear submarines, the nuclear ter areas without detection. Nuclear submarines also have fleet included four Kirov-class guided-missile cruisers, a higher speeds than their diesel counterparts, and provide small number of nuclear-powered scientific research, sup- better living conditions for their crews. Since 1952, four port, and space-tracking vessels, and seven civilian generations of nuclear submarines and several experimen- nuclear-powered . According to the Bellona tal nuclear submarines have been built for the Soviet and Foundation, a Norwegian organization which monitors Russian navies. From 1955 to 1964 a total of 55 first- naval nuclear developments in Russia, the eighth Russian generation (November-, Hotel-, and Echo- class) subma- nuclear-powered , 50th Anniversary of Victory, rines were built. At the height of the , 1 was fitted with two reactors in July 2001. After the break- approximately five to 10 nuclear submarines were being up of the Soviet Union in 1991, the financially strapped commissioned per year from each of the four Soviet sub- Russian Federation inherited the Soviet nuclear fleet. As marine yards: in ; Admiralteyskiye a result, by 1996 only 109 nuclear submarines remained Verfi in St. Petersburg; Krasnoye Sormovo in Nizhniy in service. Two-thirds of these submarines are assigned Novgorod; and Amurskiy Zavod in Komsomolsk-na-

The Nonproliferation Review/Spring 2002 161 MOHINI RAWOOL-SULLIVAN, PAUL D. MOSKOWITZ, & LUDMILA SHELENKOVA

Amure (see Figure 1). Beginning in the 1980s, the Soviet problem which has received little attention. The Soviet Union launched several titanium-hulled submarines. The Union, unlike the , built nuclear submarines ill-fated Komsomolets (Mike-class), which sank with 42 powered not only by pressurized water reactors (PWRs), crewmembers aboard in 1989, had a titanium hull. Later, but also constructed submarines with liquid metal cooled two series of nuclear submarines were constructed with (LMC) reactors. Although only eight submarines (seven titanium hulls: the Project 705 (Alfa-class) and the Project Alfa-class and one prototype) were built with LMC reac- 945 (Sierra-class). These titanium-hulled nuclear subma- tors, this report analyzes the unique proliferation and safety rines are no longer in production. The Kursk, which sank issues the decommissioning of these submarines presents. on August 12, 2000, killing all of its crew, belonged to the The report begins by discussing the development of the Oscar II-class (third generation). Third generation Akula- Russian naval propulsion program, including the origins class attack submarines and Oscar-class cruise missile of the liquid metal cooled reactor program. It then dis- submarines, fourth generation Severodvinsk-class attack cusses the design, development, production, and opera- submarines (which carry anti- cruise missiles as well tion of the Alfa-class submarines that were powered by as torpedoes), and fifth generation Borey-class SSBNs are LMC reactors. The LMC reactor fuel cycle and its pro- still in production. However, severe funding problems have liferation and safety implications are also addressed. The slowed the pace of completion and commissioning of these article concludes with an examination of the unique chal- submarines. lenges posed by the dismantling of the LMC reactors re- While the proliferation challenges posed by the disman- moved from the Alfa-class submarines, and urges tling of Russian nuclear submarines have been addressed additional international assistance to help Russian address in a number of articles, there is one unique aspect of this these challenges.

Figure 1: Locations of acilities involved in the Alfa-class/LMC reactor activities

162 The Nonproliferation Review/Spring 2002 MOHINI RAWOOL-SULLIVAN, PAUL D. MOSKOWITZ, & LUDMILA SHELENKOVA

History of Russian Russian Nuclear-Powered Submarines with LMC Table 1 shows Russian naval propulsion reactor design reactors models, their power, the uranium enrichment level of their Immediately following the successful launch of the first fuel, and the class of vessels built with these models. This Soviet nuclear-powered submarine, Project 645 was ini- table is based on data obtained from the NIS Nuclear Pro- tiated in 1957. The objective of this project was to build files Database of the Center for Nonproliferation Studies a 1,500-ton “interceptor” submarine that could get to a at the Monterey Institute of International Studies, except target location before the target information was obsolete. for the information on the number of reactors on Alfa- To meet this requirement, the submarine had to have un- class submarines.2 For many years, most Western refer- derwater speed of about 40 knots. The speed of a sub- ence material suggested that the Alfa-, Sierra-, and marine is largely a function of its wetted hull area and the Akula-class submarines were powered by two reactors. power of its engines. Thus, to obtain high speeds, a de- Recently obtained information, however, shows that these sign with a smaller volume hull and high-powered engines submarines were powered by just one reactor.3 is required. A. B. Petrov of the Malakhit Design Bureau Russian research into naval reactors began in the early (SKB-143) in Leningrad came up with an innovative de- 4 1950s. Veterans of these early programs take great pride sign for such a submarine. A LMC reactor design was in these efforts because of the originality of their programs, selected to provide high power using a compact reactor. which were not copied from U.S. or other designs. Two The LMC reactor shielding consisted of a single wall in- options were considered from the early stages of devel- side the submarine, although normally single-wall shield- opment. The first approach used a water-cooled, water- ing would be considered inadequate. The engine plant was moderated design (also known as a PWR) developed at completely automated and expected to run unmanned. The the Kurchatov Institute in Moscow under the leadership crew in this first design was expected to be 15 to 17 per- of A. Alexandrov. The second approach used a lead-bis- sonnel. The complexity of this design required that all muth cooled reactor design (also known as an LMC reac- crewmembers be officers. If this design were built with a tor) developed at the Institute of Physics and Power steel hull, however, the weight of the submarine would Engineering in Obninsk under the leadership of A. not allow it to meet the required performance specifica- Leipunsky. Over the next 40 years, the Soviet Union de- tions. This problem was resolved by using a titanium al- veloped and produced three generations of naval PWRs. loy hull. The use of titanium alloy allowed the hull walls Each generation featured improved reliability, compact- to be thinner and reduced overall weight. Around 1963, ness, and quietness. The first generation PWRs (VM-A after many delays, Petrov was dismissed from the project, type) were deployed from 1957 to 1968. All of these re- and under the direction of M. G. Rusanov, the design of actors are now retired. The second generation, VM-4 type reactors were deployed from 1968-1987. As of 1995, some of these reactors were still in use. The third generation Figure 2: Alfa-class submarine reactors (OK-650 type) entered service starting in 1987. Besides these reactor designs, several one-of-a-kind PWR designs were developed for research submarines and ves- sels. Despite certain advantages gained by using heavy metal- coolant in a submarine reactor, the LMC reactor design was ultimately abandoned in favor of the PWR designs. This decision was motived by safety considerations, main- tenance problems, accidents, and advances in PWR de- signs. The main difficulty with LMC reactors arose from the need to keep these reactors sufficiently hot while the submarine was in port. If the reactor temperature fell be- low about 123 degrees Celsius (oC), the liquid-bismuth coolant would congeal, causing the reactor to seize up and essentially “freeze” itself. Source: Federation of American Scientists

The Nonproliferation Review/Spring 2002 163 MOHINI RAWOOL-SULLIVAN, PAUL D. MOSKOWITZ, & LUDMILA SHELENKOVA

the interceptor submarine was altered. The new design compartments: (1) room; (2) accumulators and increased the displacement to 2,300 tons and doubled the living accommodation; (3) control room; (4) reactor com- number of internal compartments.5 partment; (5) turbo and diesel generators, cooling and This design was ready for construction in 1965. The auxiliary machinery; (6) turbines; and (7) generator. The first submarine to have LMC reactors was designated K- main purpose of K-27 was to test the functioning of LMC 27, and used a Project 627 ZhTS-class hull. This vessel reactors on a submarine. The main electrical system was was designed by Special Design Bureau Number 143 designed to operate at a frequency of 400 Hertz, which in (SKB-143) in Leningrad and was constructed at the turn permitted a reduction in size of some of the reactor Severodvinsk . The hull was divided into seven equipment. For example, the capacity of the batteries in

Table 1: Naval Propulsion Reactor Design

FUEL ENRICHMENT MODEL POWER (% Uranium-235) VESSELS 2 PWR*/VM-A 70 MWt 20% Hotel, Echo, November

70-90 1PWR/VM-4 20% Charlie MWt

70-90 2PWR/VM-4 20% Victor II and III, Delta, Yankee MWt

2 PWR/OK-650 190 MWt 20-45% Typhoon, Oscar

1 PWR/OK-650 190 MWt 20-45% Akula, Sierra, Mike

2PWR/VM-5m 177MWt unknown Papa 2 LMR**/VT-1 73 MWt weapons-grade November 645

2LMR/OK-550or 155 MWt weapons-grade Alfa BM-40A

10 (X- 1 PWR/type unknown unknown X-Ray, Uniform, AS-12 Ray)

2 PWR/KN-3 300 MWt unknown Kirov (battle cruiser)

2 PWR/type unknown 171 MWt unknown Kapusta (auxiliary ship)

2-3 PWR/OK-150 and 90 MWt 5% Lenin (icebreaker) OK-900

2 PWR/KLT-40 135 MWt up to 90% Arktika (icebreaker)

Sevmorput (auxiliary ship), 1PWR/KLT-40 135MWt upto90% Taymyr (icebreaker)

Source: “Russia: Naval Nuclear Reactors,” Center for Nonproliferation Studies, NIS Nuclear Profiles Database *PWR is pressurized water reactor. **LMR is liquid metal reactor. In this report, “weapons-grade uranium” refers to uranium enriched to 90 percent or more U-235.

164 The Nonproliferation Review/Spring 2002 MOHINI RAWOOL-SULLIVAN, PAUL D. MOSKOWITZ, & LUDMILA SHELENKOVA

K-27 was only one-fourth that in submarines equipped this situation, the hazard came from radioactive aerosols with PWRs. The submarine was also equipped with au- and gaseous polonium compounds (combined with lead tomatic turbo generators. and in some cases with hydroxide). Polonium-210, which K-27 was equipped with two VT-1 type LMC reac- is an alpha emitter with a lifetime of 13.8 days, is gener- tors. But the submarine suffered a major accident with ated by the interaction of neutrons (generated in the reac- one of its reactors in 1968, and nine members of the crew tor core) with bismuth-209 from the lead-bismuth coolant 8 died. In this accident, one of the reactors of the subma- (about 44 percent lead and 56 percent bismuth). This rine was severely damaged after a pipe in the reactor com- reaction generates bismuth-210, which decays into Po- partment was contaminated by corrosion particles from 210. After emission of an alpha particle, polonium-210 the liquid metal coolant. After the accident, the core was decays into stable lead-206. These problems were ana- removed from the undamaged reactor. All empty spaces lyzed and technical measures were developed to prevent 9 in the damaged reactor compartment including the empty their recurrence. The main technical parameters of the cavities in the reactor itself were filled with furfural and land prototype LMC reactor are listed in Appendix A, Table bitumen developed by the Kurchatov Institute. In 1981, A-1. Table A-2 lists the characteristics of the core of the the entire vessel was scuttled in the Kara Sea, near Novaya land prototype LMC reactor of facility 27/VT. In 1978, Zemlya. another full-scale prototype of the LMC reactor was put into operation at the KM-1 facility at the A.P. Aleksandrov Despite this accident, the experience gained with Project Scientific Research Technological Institute (NITI), in 645 led to the design and construction of seven Project Sosnovy Bor.10 705 and 705 K-class or Alfa-class submarines. All of these submarines were equipped with LMC reactors. The Alfa- The reactors of the Alfa-class submarines were never class submarines were built for speed (40+ knots); hence refueled, for it was simply not technically possible to re- it was of small consequence that they were noisy, as it move the fuel assemblies without the metal coolant so- was assumed they could escape from any torpedoes fired lidifying in the process. It is true that a team led by B. at them. Since Alfa submarines were designed to be “in- Gromov published an article in which they report that they terceptor” submarines, their operational endurance was froze the undamaged core of K-27, kept it in this state for relatively short—approximately 50 days. The first Alfa- two years, and then thawed the reactor core and brought 11 class submarine experienced numerous problems, and the it up to high power. However, this success must have last Alfa-class submarine, K-123, was decommissioned been achieved under controlled experimental conditions, around 1995. All of the Alfa submarines were lightly armed. and the procedure could not be used for routine opera- Equipment in these submarines was unique, and thus hard tional refueling. As a result, Alfa-class submarines were to maintain. never refueled once the coolant in their reactors was fro- zen. Liquid Metal Cooled Reactors Since the mixture of lead and bismuth utilized in the LMC reactor has a high boiling point (1670°C ) and a In 1953, the Soviet Union began construction of a land- melting point of 123°C, it is unnecessary to keep LMC based full-scale prototype of a LMC reactor and associ- reactors under pressure as is the case for the pressurized ated propulsion unit, called a 27/VT. This facility was water-cooled reactors.12 Conversely, it is important to keep commissioned at Obninsk in January 1959.6 It became LMC reactors constantly heated above 123°C so that the the research and development base and training facility metal solution does not solidify. Once the solution hard- for naval LMC reactors. The facility operated for two run ened, it was impossible to restart these reactors, as the periods lasting a total of 17 years. Two different cores fuel assemblies were then frozen in the solidified coolant. were used for these runs. After the second run, research- A special land-based facility was constructed at Zapadnaya ers noticed a high content of slag in the reactor core.7 In Litsa, the base of the Project 705–705 K- or Alfa-class these tests they also noticed the coolant freezing prob- submarines, to deliver superheated steam to the LMC re- lems in some sections of lead-bismuth circuits due to ei- actors of the moored vessels. In addition, a smaller ship ther emergency situations or personnel errors. and floating barracks were also stationed at the pier to Yet another problem arose when the primary coolant deliver steam from their own boilers to the Alfa subma- circuit sprung leaks in the course of scheduled repairs. In rines at the piers. This method of providing external heat-

The Nonproliferation Review/Spring 2002 165 MOHINI RAWOOL-SULLIVAN, PAUL D. MOSKOWITZ, & LUDMILA SHELENKOVA

ing proved to be unreliable, suffering breakdowns in the penetrated by 10 control and compensation rods (CCRs) 1980s. After this failure, the reactors of all operational Alfa- and 3 emergency protection rods. class submarines were kept running continuously until they Another source of information on LMC reactor fuel is were decommissioned in the 1990s. Consequently, while data collected during Project Sapphire.16 Project Sapphire, the reactors were running, the vessels needed to be carried out in 1994, involved the packaging and transfer manned. of 581 kg (1,278 lbs) of highly (HEU) In theory, the following factors were expected to make from the Ulba Metallurgical Plant near Ust’-Kamenogorsk the LMC reactor design safer overall than the PWR de- in Kazakhstan to the Y-12 Plant at the Oak Ridge Na- sign:13 tional Laboratory in Tennessee. 17 The project was initi- • The high boiling point of the metal coolant mixture ated by President Nursultan Nazarbayev of Kazakhstan, (approximately 1680°C) means that there is no possi- in consultation with Russia, in order to prevent the pos- bility of over-pressurization of the primary circuit in the sible theft of this material by terrorists or proliferants states. event of accidental overheating of the coolant. The HEU involved was reportedly left over from the So- • Since it is impossible to overheat the coolant, heat viet Alfa submarine program, and had been stored at Ulba removal from the core is more reliable, increasing safety. in unsecured and unsafeguarded facilities, without elec- • Reactor power decreases in the event of both emer- tronic means of accounting.18 Instead, material invento- gency overheating and the simultaneous failure of the ries were simply recorded by hand into books. The emergency protection system. material was in seven forms: HEU metal (168.7 kg); HEU • In the core and within the coolant there is no release oxides (29.7 kg); beryllium-HEU alloy fuel rods (148.6 of hydrogen from irradiation or chemical reactions. The kg); beryllium-HEU alloy machining scrap and powder coolant itself also reacts only very slightly with air or (231.5 kg); beryllium oxide-uranium dioxide fuel rods (1.6 water. As a result, there is no danger of an explosion in kg); graphite with trace HEU (0.7 kg); and laboratory sal- the event of a coolant leak. vage (0.2 kg).19 The average enrichment of the material • If a leak did take place, the coolant would rapidly was 89-90 percent U-235. 20 The LMC reactor of each solidify. The melting point of the coolant mixture is Alfa class submarine is believed to have contained approxi- 123°C, thereby removing the possibility of reactor dam- mately 200 kg of HEU.21 age and loss of coolant. Two different models of LMC reactors were developed. Detailed information on the LMC reactors installed on The four submarines built at the Admiralty Shipyard used the Alfa-class submarines is hard to find. In their two pa- the BM-40A reactor with two separate steam loops and pers, “Possible Criticality of the Marine Reactors Dumped circulating pumps. The submarines built at Severodvinsk in Kara Sea,”14 and “Potential Radionuclide Release Rates (project 705K) used the OK-550 with branched first-loop from Marine Reactors Dumped in the Kara Sea,”15 J. M. lines and triple circulating loops and pumps. Both models Warden and his colleagues describe the LMC reactors. were equally susceptible to leaks and meltdowns. According to these papers, the LMC reactor of the sunken November class submarine (K-27) consisted of a cylin- The LMC Reactor Fuel Cycle and Proliferation drical stainless steel (SS) reactor pressure vessel with the Implications following dimensions: 1.8 meters (m) in diameter, 3.7 m Initially, the HEU-beryllium alloy fuel for the LMC re- in height, and 30 millimeter (mm) thick walls. These LMC actors was produced at the Ulba Metallurgical Plant. Pro- reactor cores were loaded with 90 kg of uranium-235 en- duction of this fuel at Ulba ended in the 1970s and the riched to 90 percent and clad in SS. The core was made production of all naval reactor fuel was consolidated at up of a triangular lattice of ceramic fuel pins composed of the Machine Building Plant in Elektrostal near Moscow.22 highly enriched uranium alloyed with beryllium and sin- From Elektrostal, fuel was delivered to the Navy, which tered with beryllium oxide. The pins were surrounded by conducted most refueling. Fresh fuel was also delivered lead-bismuth coolant. The core radius was 390 mm and to submarine construction for fuelling newly- was surrounded by a stainless steel (SS) layer and a be- built submarines and for refueling submarines undergoing ryllium oxide reflector, with more lead-bismuth coolant major overhauls. The spent fuel from these submarines and SS situated above and below the core. The core was was to be transported to the Mayak Production Associa- tion in Chelyabinsk Oblast, Russia, for reprocessing.

166 The Nonproliferation Review/Spring 2002 MOHINI RAWOOL-SULLIVAN, PAUL D. MOSKOWITZ, & LUDMILA SHELENKOVA

LMC reactor cores are loaded or unloaded in the form sian naval bases have frequently failed to receive their of a removable unit referred to as removable reactor core wages for months, increasing the risk of sabotage or unit (RRCU).23 Each RRCU includes the core with fully theft.29 inserted neutron-absorbing emergency protection rods There have been several documented cases involving 24 (EPRs), reflector, and some shielding material. The re- the theft of naval propulsion reactor fuel. A sailor and a moval of the RRCU can be performed either with or with- guard were arrested in July 1993 while attempting to steal out coolant in the reactor vessel. When a spent RRCU is two fuel rods containing 1.8 kg of HEU (enriched to 36 removed from a reactor vessel, its EPRs are left in fully percent U-235) from a storage site at the Zapadnaya Litsa inserted positions. The driving mechanisms of these EPRs naval base. In another case, two naval officers diverted are then dismantled and SS caps are welded onto the EPR 4.5 kg of HEU (20 percent U-235) from Sevmorput ship- 25 inserts. This procedure aims to prevent any accidental yard in November 1993. Although these and other docu- removal of the EPRs from a spent RRCU. These spent mented cases have not involved fuel from Alfa-class RRCUs are put into a steel container with non-irradiated submarines, the threat of theft remains high for the LMC lead-bismuth eutectic at a temperature ranging from about RRCUs. It is notable that most of the cases of naval pro- 150°C to 160°C. After the spent RRCU has been placed pulsion reactor fuel theft have involved insiders. In addi- in the steel storage container the temperature of the lead- tion, there are also reports of arrests of Chinese and North bismuth eutectic is allowed to fall below its melting point. Korean nationals on Russian submarine bases.30 As a result, spent RRCUs are stored in the frozen lead- bismuth eutectic. The steel containers holding the spent Decommissioning and Dismantling of the Alfa-Class RRCUs are then sealed and a layer of bitumen applied to prevent moisture penetration. The maximum estimated Submarines heat release in loaded or unloaded RRCUs is 3 kW.26 All Decommissioning (removal from active service) and of the spent RRCUs from Alfa-class submarines are stored dismantling (physical breakdown) of Alfa-class submarines at the Gremikha naval base in Oblast, Rus- is carried out in several stages. First, all weapons and ex- sia.27 The rest of the Alfa-class submarines have fuel on- plosives are removed, along with classified and sensitive board with frozen coolant. The procedures described equipment. Next the submarine is taken to a facility above for the storage of LMC RRCUs were intended for equipped to dismantle it. At this facility, the submarine’s short-term storage (three to five years) and not for long- reactors are shut down so that defuelling can take place. term storage. Some of the LMC RRCUs have been at At this time, additional equipment is removed from the Gremikha for over 10 years.28 It is possible that storing submarine, such as loose furnishings, expendable materi- RRCUs over long period of time in short-term storage als, tools, and spare parts. Before defuelling begins the could cause safety and/or security problems as a result of submarine is dry-docked. The hull above the reactor plant the impact of physical and chemical processes (galvanic is opened. Special equipment necessary for the removal and chemical corrosion, phase conversion). RRCUs in of the reactor core is installed. Removal and recovery of short-term storage may also be vulnerable to possible theft, reusable materials from the reactor takes place. Then the terrorist attack (sabotage), earthquakes, and fires. fuel is then removed and stored in short-term storage area Physical protection of fresh or spent naval propulsion or transported to an appropriate facility for reprocessing. reactor fuel is remains a matter of serious concern owing Defuelling removes over 99 percent of the radioactivity to inadequate safety and security measures at many Rus- associated with the reactor. Once the fuel has been re- sian storage facilities. The spent fuel from Alfa-class sub- moved, most residual radioactivity (about 99 percent) in marines still contains a large amount of (HEU). It is worth the reactor compartment is concentrated in the reactor noting that separation of HEU from low-irradiated spent vessel structures and the primary circuit pipelines. This fuel is much easier than chemical reprocessing required residual radioactivity is caused by neutron activation of for plutonium separation and can be done at smaller fa- the reactor steel structure and radioactive deposits of cool- cilities. As a result, spent naval fuel would be attractive to ant residue on the inner surfaces of the primary circuit terrorist groups or proliferants states seeking a relatively pipes. After defuelling is completed, the reactor compart- easy route to obtaining the necessary fissile material to ment is cut away from the rest of the hull. The reactor fabricate a nuclear weapon. In addition, workers at Rus- compartment and the remaining reactor parts are then packaged for long-term storage. The remainder of the

The Nonproliferation Review/Spring 2002 167 MOHINI RAWOOL-SULLIVAN, PAUL D. MOSKOWITZ, & LUDMILA SHELENKOVA submarine is scrapped. Decommissioning a submarine is tional sections to its stern and nose ends for buoyancy a labor-intensive task, and requires involvement of the support, was shipped to the sediment stop. Probably, submarine crew for maintenance, reactor monitoring and in the future the fuel can be unloaded from the removed defuelling purposes. reactor. It took 8-9 years to change the reactor com- Alfa-class submarines are dismantled at the Sevmash partment and the submarine was finally launched again shipyard in Severodvinsk, located in . in 1990 (renumbered B-123). Recommissioned in 1991, Officially, the decommissioning of nuclear submarines is it remained in operation till 1993. It was scheduled for supposed to be self-financing. The profits from the sale decommissioning during 1995. At the start of the de- of scrap material are deemed adequate to fund the entire commissioning the coolant was frozen in the reactor decommissioning process.31 In reality, however, the ship- and NS was shipped afloat to the sediment stop. yards that carry out submarine dismantlement suffer ma- • K-432: (Project 705K) This submarine is in the pro- jor financial losses. These financial losses are even greater cess of being dismantled at the Sevmash shipyard. In a when the titanium-hulled submarines—such as Alfa-class freak accident, the submarine hit a whale during its sea submarines—are dismantled.32 Greater losses are incurred trials, necessitating major repairs that were completed 37 in dismantling these submarines because dealing with the in 1988. However, the submarine was never re-com- titanium hull requires more time and advanced equipment. missioned. The LMC RRCU from this submarine was The management of Sevmash estimates that the shipyard removed in Gremikha and is being stored there. Its re- lost one billion rubles (about $2.5 M) on the decommis- actor compartment was towed to Sayda Bay on the 38 sioning of 705 - Alfa class submarines K-463.33 Kola Peninsula for storage in 1998. • K-463: This submarine had unspecified reactor acci- dent. It was decommissioned at Sevmash sometime Current Status of Alfa-class submarines after 1986. The LMC RRCU has been removed from The current status of the seven Project 705 and 705 the reactor and is being stored at Gremikha. This sub- K–class or Alfa-class submarines is summarized below:34 marine was later re-engineered with standard VM-4 • K-377, (formerly K-47): (Commanding Officer: A.S. reactor (Project 671B) and served as a trials ship. In Pushkin) This was the first Alfa class submarine. It 1994, the reactor compartment was towed to Sayda suffered a reactor accident in 1972 during its sea trials. Bay, and remains moored there today. The remainder The metal coolant “froze” and it became impossible to of the submarine was scrapped. remove the reactor fuel. After this accident, the sub- • K-493:39 (Project 705K) The LMC RRCU was re- marine was dismantled and the whole reactor compart- moved from this submarine at Gremikha sometime ment (with the LMC reactor inside) was filled with before 1995 and is now stored there. The submarine furfural and bitumen.35 This reactor compartment was was then reengineered with VM-4 (project 671B) and then placed on a barge for transport to the Kara Sea served as a training vessel. Subsequently, the subma- where it was to be dumped. However, the signature rine was stationed at the Zapadnaya Litsa (Bolshaya by the Soviet Union of the 1972 London Convention Lopatka) naval base on the Kola Peninsula untill 1996. prevented the implementation of this plan. Subse- It was then towed to the Sevmash shipyard for disman- quently, the barge was towed to the island Yagry out- tling, which was completed in November 1997. side the Zvezdochka shipyard in Severodvinsk, where • K-373: This submarine is laid up at Zapadnaya Litsa it was stored. Sometime in late 1994 or early 1995 the and has not yet been defuelled. At the start of its de- reactor compartment was moved to Gremikha for stor- commissioning a problem occurred at the reactor lid age on shore. due to destruction of the EPR. Some radioactive dust • K-123: Built at Severodvinsk (Project 705K).36 penetrated into the upper space of the EPR inserts (in Launched on December 26, 1977. The original reactor the reactor lid area), resulting in the deterioration of the compartment was removed in 1982 following an acci- radiation situation in the reactor compartment. At present dent, and a new one installed. Liquid metal from the the coolant in the reactor is frozen primary cooling circuit leaked out and contaminated the • K-316: Dismantling of this submarine began in the entire reactor compartment. The contaminated reactor autumn of 1994 at the Sevmash shipyard. The LMC compartment was cut out and, after welding the addi- RRCU was removed at Gremikha and is currently stored there. The reactor compartment was towed to

168 The Nonproliferation Review/Spring 2002 MOHINI RAWOOL-SULLIVAN, PAUL D. MOSKOWITZ, & LUDMILA SHELENKOVA

Sayda Bay in 1995. The rest of the submarine was United States. Although the U.S. Department of Defense scrapped. Cooperative Threat Reduction (CTR) program is financ- In addition to the RRCUs noted in the above list, there ing the dismantlement of Russian ballistic missile subma- are also two spent RRCUs from the first run of the Project rines, as well as the storage or reprocessing of naval 645 stored at the Gremikha storage propulsion reactor fuel removed from them, the mandate facility. Fuel elements from the two cores of the Obninsk of this program does not extend to non-strategic weapons 27/VT prototype facility have been removed from the systems, including attack submarines like the Alfa class. RRCUs there, and both the fuel elements and the remain- Beyond the proliferation threat, it is also important to ing RRCU parts are currently stored on-site at Obninsk. consider the criticality safety of RRCUs removed from the Alfa-class submarines. Current storage, both land- Future Challenges based storage and on board submarines awaiting defuelling, was intended to be short-term. Safe long-term Storage of the RRCUs of Alfa-class submarines raises storage procedures for the Alfa-class RRCUs need to be a number of difficult issues. Owing to the high enrich- developed and analyzed. At this point it is hard to evalu- ment level and large quantity of fuel used in LMC reac- ate the criticality safety issues related to the stored RRCUs tors, there is a significant proliferation threat if the RRCUs in any depth because of inadequate technical data and are not adequately secured. The spent fuel from LMC analysis. To understand the criticality issues posed by us- reactors could be attractive to terrorists and proliferant ing short-term storage facilities for long-term storage, the states seeking fissile material for nuclear weapons. At great- physical and chemical properties of lead-bismuth coolant est risk of theft are spent and partially spent fuel elements over a long period of time while in a frozen state must be separated from the RRCUs. By contrast, fuel elements better understood. The corrosive and other effects of frozen that have not yet been removed from the frozen RRCUs coolant (if any) on SS and other materials used to store are less vulnerable to theft, because substantial special- the RRCUs must also be studied. A better knowledge of ized equipment is required to move a whole RRCU. The these processes and the details of LMC reactor design risk of theft is very low for RRCUs that are still frozen in would facilitate an accurate evaluation of the risks of a lead-bismuth coolant. However, if at some point these fuel criticality incident arising under in various scenarios, such elements are unloaded from their RRCUs, then adequate as the leakage of water into stored RRCUs. physical protection must be provided. At the moment, important details about the LMC RRCUs As noted above, spent fuel elements from the proto- used on the Alfa-class submarines are unavailable, and type facilities are currently stored at the Institute of Phys- remain classified. These details are needed to calculate ics and Power Engineering in Obninsk. It is logical to criticality risks. Other issues such as fires at the storage assume that there are also stocks of fresh fuel stored some- facilities or on submarines must be studied. The effects where in Russia. A careful evaluation of the physical se- of possible earthquakes on stored RRCUs should also be curity of all spent and fresh fuel elements for LMC reactors evaluated. Given a criticality accident, a consequence analy- is needed. The reactors of all Alfa-class submarines with sis may be performed to determine the environmental fuel on board are apparently frozen as a result of various impact. This would include an analysis of the accident accidents and technical problems. These reactors are not transient to determine the physical state of the fuel after operable, and as a result, it would be virtually impossible the accident and the additional fission product inventory for a terrorist group of proliferants state to attempt to steal generated. The results of this analysis may be applied to one of these Alfa-class submarines. Still, c urrent politi- a dispersion analysis to track the release of radioactive cal and economical conditions in Russia make it clear that materials into the environment. The international commu- in order to further reduce proliferation risks, the interna- nity should provide technical assistance and monetary tional community should provide technical assistance and support for such studies. For its part, if it wants to see funding to ensure adequate security of the spent and fresh both the security and safety issues associated with the dis- fuel elements from the LMC reactors and the dismantle- mantling of the Alfa-class submarines adequately ad- ment of those Alfa-class submarines that still await dressed, Russia must be willing to increase cooperation in defuelling. Meeting this challenge may require new legis- this area and share design-specific knowledge with inter- lation and new thinking in donor countries, such as the national partners.

The Nonproliferation Review/Spring 2002 169 MOHINI RAWOOL-SULLIVAN, PAUL D. MOSKOWITZ, & LUDMILA SHELENKOVA

In summary, both the proliferation threat and safety cooperation. Negotiations are currently underway to es- issues posed by the dismantling of Alfa-class submarines tablish international cooperation in the dismantling of Rus- and the storage of their spent and fresh fuel should be sian attack submarines. These talks are both bilateral (e.g., regarded as global issues, and not just solely as a Russian Norway-Russia and Japan-Russia) and multilateral (e.g., concern. In our view, the ideal scenario would be interna- the International Atomic Energy Agency Contact Expert tional cooperation that ensured that the Alfa-class subma- Group for International Spent and Radioac- rines are dismantled and spent fuel is either stored in safe tive Waste Management). The success of these negotia- and secure storage or is reprocessed. To achieve this ideal tions will depend, in addition to those issues noted above, scenario, negotiations between Russian and the interna- on Russia agreeing to sign international nuclear liability tional community must hammer out a basis for expanded indemnification agreements.

TECHNICAL APPENDIX A

Table A-1: Technical parameters of the ground prototype LMC reactor at the 27/VT facility

Parameter Value Reactor Thermal Power, kW 70,000 Coolant Flow Rate, m3/h 850 Eutectic Temperature at the Reactor Inlet, oC 235 Eutectic Temperature at the Reactor Outlet, oC 440 Steam Generation (SG) Capacity, t/h 90 Pressure of superheated steam at the SG outlet, 38 kg/cm3 Temperature of the Superheated Steam, oC 385

Table A-2: Core parameters of the ground prototype LMC reactor at the 27/VT facility

Parameter Value Core Diameter 769 mm Core Height 853 mm Content of U in U-Be alloy 7-16 % Diameter of U-Be core 11 mm Triangular Lattice Pitch 13.6 mm Number of rod fuel elements 2735 The number of control and safety system (CSS) rods 16 (absorber is natural boron carbide)

Sources: Suvorov, et al., “Experience of 27/VT facility construction and operation,” Proceedings of the International Conference on Heavy Liquid Metal Coolants in Nuclear Technology, Vol. 1 (Obninsk, Russia: State Science Center of the Russian Federation-Institute of Physics and Power Engineering, 1999), p. 67.

170 The Nonproliferation Review/Spring 2002 MOHINI RAWOOL-SULLIVAN, PAUL D. MOSKOWITZ, & LUDMILA SHELENKOVA

1 “200t Reactor Installed On Board Icebreaker,” July 27, 2001, Bellona Founda- 21 Oleg Bukharin, “Analysis of the Size and Quality of Uranium Inventories in tion Web Site, . Russia,” Science and Global Security 6 (1996), pp. 59-77. 2 “Russia: Naval Nuclear Reactors,” Monterey Institute of International Stud- 22 Oleg Bukharin and Joshua Handler, “Russian Nuclear-Powered Submarine ies, Center for Nonproliferation Studies, NIS Nuclear Profiles Database, Decommissioning,” Science and Global Security 5 (1995), pp. 245-271. . 23 V.K.Sazonov, V.I.Tihomirov, V.B.Tihonenko, G.I. Toshinsky, A.I. Shestopalov, 3 Thomas B. Cochran, Robert S. Norris and Oleg A. Bukharin, Making the V.S. Stepanov, M.P. Vahrooshin, and G.A. Alexeev, “Technology and Experi- Russian Bomb: From Stalin to Yelstin (Boulder, CO: Westview Press, 1995), ence of Repair Works and Refueling for the Nuclear Power Installations with p. 287. Lead-Bismuth Coolant,” Proceedings of the International Conference on 4 Anatoliy Valeryevich Kuteynikov, “Malakhit Design Bureau: Creator of the Heavy Liquid Metal Coolants in Nuclear Technology, Vol. 2 (Obninsk, Rus- First Soviet Nuclear Powered Submarine,” Vooruzheniye, Politika, sia: State Science Center of the Russian Federation-Institute of Physics and Konversiya, No. 2 (1994), pp. 18-21, . 24 Ignatiev Sviatoslav, Dmitry Pankratov, Boris Gromov, Evgueni, Victor 5 “K.377 Class Nuclear Attack Submarines,” Warships1.com website, Kalchenko, and Vladimir Stepanov, “The Safety Assessment During the Long- . Term Storage of Spent Fuel of Alfa-Class Submarines,” poster presentation I- 6 G.P. Suvorov, O.V. Kuzko, and M.I. Bugreev, “Experience of 27/VT Facility P17 at the ECO-INFORMA 2001 Conference, Argonne, IL May 14-18, 2001. Construction and Operation,” Proceedings of the International Conference 25 Mikhail Bugreev, Boris Gromov, Evgueni Efimov, Sviatoslav Ignatiev, Victor on Heavy Liquid Metal Coolants in Nuclear Technology, Vol. 1 (Obninsk, Kalchenko, Dmitry Pankratov, Valery Sazonov, and Mikhail Vakchrushin, “Chal- Russia: State Science Center of the Russian Federation-Institute of Physics lenges for Long-Term Storage of Spent Fuel of Alfa-Class Nuclear Subma- and Power Engineering, 1999), p. 67.. rines,” paper presented at ECO-INFORMA 2001 Conference, Argonne, IL, 7 B.F. Gromov, G.I. Toshinsky, V.V. Checkunov, Yu.I. Orlov, Yu.S. Belomytsev, May 14-18, 2001. I.N. Gorelov, A.G. Karabash, M.P. Leonchuk, D.V. Pankratov, and Yu.G. 26 Ibid. Pashkin, “Designing the Reactor Installation with Lead-Bismuth Coolant for 27 Niels Boehmer, Aleksandr Nikitin, Igor Nilsen, Michael H. Mcgovern, and Nuclear Submarines,” Proceedings of International Conference on Heavy Andrey Zolotkov, The Arctic Nuclear Challenge, Bellona Report Volume 3 Liquid Metal Coolants in Nuclear Technology, Vol. 1 (Obninsk, Russia: State (Oslo: Norway, Bellona Foundation, 2001), . gineering, 1999), p. 14. 28 Mikhail Bugreev, et al, “Challenges for Long-Term Storage of Spent Fuel.” 8 B.F. Gromov, Yu.S. Belomitcev, E.I. Yefimov, M.P. Leonchuk, P.N. Martinov, 29 Igor Kudrik and Alexey Kilimov, “Biggest Russian Submarine Builder Closed Yu.I. Orlov, D.V. Pankratov, Yu.G. Pashkin, G.I. Toshinsky, V.V. Chekunov, B.A. Down… Temporarily,” Bellona Foundation Web Site, September 1998, Shmatko, and V.S. Stepanov, “Use of Lead-Bismuth Coolant in Nuclear Reac- . tors and Accelerator Driven Systems,” Nuclear Engineering and Design 173 30 James Clay Moltz and Tamara C. Robinson, “Dismantling Russia’s Nuclear (1997), pp. 207-217; O. Schipakin, N. Borisov, and S. Churkin, “Experimental Subs: New Challenges to Non-Proliferation,” Arms Control Today 29 (June Study of Po-210 Release from 17LI-83Pb Eutectic,” Fusion Engineering and 1999), . Design 29 (1995), pp. 164-169. 31 Thomas Nielson, Igor Kudrik and and Alexander Nikitin, The Russian North- 9 B.F. Gromov, et al., “Use of Lead-Bismuth Coolant.” ern Fleet, Bellona Report Volume 2, (Oslo, Norway: Bellona Foundation, 1996), 10 Ibid. . 11 Ibid. 32 Some sources believe that the last two Alfa submarines might have been 12 Ibid. built using SS instead of titanium alloy. References such as Boehmer, et al., 13 Ibid. The Arctic Nuclear Challenge, report that all Alfa-class submarines were built 14 J. M. Warden, M. Mount, and N. M. Lynn, “Possible Criticality of the using titanium alloy. This source also says that the BM-40A reactor was in- Marine Reactors Dumped in Kara Sea,” paper presented at the Third Interna- stalled on all Alfa-class submarines. tional Conference on Environmental Radioactivity in the Arctic, Tromso, Nor- 33 Thomas Nielson, et al., The Russian Northern Fleet. way, June 1-5, 1997, UCRL-JC-125452. 34 Bugreev, et al., “Challenges for Long-Term Storage of Spent Fuel;” and 15 J. M. Warden, N.M. Lynn, M. Mount, Y. Sivintsev, S.J. Timms, E. Yefimov, K. Boehmer, et al., The Arctic Nuclear Challenge. Gussgard, R. S. Dyer, and K-L. Sjoeblom, “Potential Radionuclide Release 35 John H. Heiser, Yuri Sivintsev, Valeri. P. Alexandrov and Robert. S. Dyer, Rates from Marine Reactors Dumped in the Kara Sea,” The Science of the “Leaching of Radionuclides from Furfural-based Polymers Used to Solidify Total Environment 202 (1997), pp. 225–236. Reactor Compartments and Components Disposed of in the Arctic Kara Sea,” 16 R. G. Taylor, “Description of Project Sapphire,” paper Submitted to the Fifth paper presented at RADOC 1999 conference, Edinburgh, Scotland, September International Conference on Nuclear Safety, Albuquerque, NM, September 17- 1999; and John H. Heiser, Melvin G. Cowgill, Yuri Sivintsev, Valeri P. Alexandrov, 22, 1995, Y/DD-667/R1 and Robert. S. Dyer, “Furfurol-Based Polymers for the Sealing of Reactor 17 According to some news reports in 1996, this total amount of 581 kg was not Vessels Dumped in the Arctic Kara Sea,” paper presented at the International accurate. Symposium on Radionuclides in the Oceans, RADOC 1996 conference, Octo- 18 “Kazakhstan: Project Sapphire,” Monterey Institute of International Stud- ber 7-11, 1996, Cherbough-Octeville, France. ies, Center for Nonproliferation Studies, NIS Nuclear Profiles Database, 36 “: Project 705 Lira Alfa-class (Nuclear Pow- http://cns.miis.edu/db/nisprofs/kazakst/fissmat/cfissmat/sapphire.htm ered),” Federation of American Scientists Web Site, . CA and Washington, DC: Monterey Institute of International Studies and 37 “Specifications: K.337 Class Nuclear Attack Submarines,” Warships1 Web Carnegie Endowment for International Peace, March 1998). Site, . 20 S. N. Cramer, K. D. Lewis, and S. D. Moses, “Project Sapphire Uranium- 38 Boehmer, et al., The Arctic Nuclear Challenge, pp. 98-99. Beryllium Dose Rate Analysis,” paper presented at the Health Physics Society 39 Igor Kudrik, “Fourth Alfa Class Sub Scrapped,” Bellona Foundation Web 1995 Annual Meeting, Boston, MA. June 23-27, 1995, YLF-516. Site, November 18, 1997, .

The Nonproliferation Review/Spring 2002 171